[foaf-dev] Expressing twitter account name in triples

Leif Warner abimelech at gmail.com
Fri Jun 17 20:12:09 CEST 2011


When I first saw the FOAF spec, I was turned off by all the arbitrary old
service-specific properties like ICQid or whatever.  But now, I would
heartily second Alexander's suggestion.  The shorthand needn't be mutually
exclusive with the long form.  With any kind of reasoning, it's trivial to
define a rule such as:

account(?person, ?blah), accountServiceHomepage(?blah, <http://twitter.com>),
accountName(?blah, ?id) <- twitterID(?person, ?id)

You may want a more general rule, so you can just say twitterID represents
id of an OnlineAccount with a homepage of "http://twitter.com".

It's nicer and less verbose to work with one triple instead of three, and
let the reasoning take care of the rest.

My vote would either be to add a shorthand like that every time someone
cares enough about it to request it, OR, get rid of all the rest of those
site-specific properties.
-Leif Warner

On Fri, Jun 17, 2011 at 7:14 AM, Bob Ferris <zazi at elbklang.net> wrote:

> Hello everybody,
>
> On 6/17/2011 3:33 PM, Michael Haschke wrote:
> > Hello,
> >
> > 2011/6/17 Alexander Gödde<alexander.goedde at tavendo.de>
> >
> >> First of all: when is “foaf:twitterID” going to be added to the
> vocabulary? The model for this is clear from predicates such as
> “foaf:skypeID”, and twitter by now has reached an importance comparable to
> the chat or messaging services that predicates are provided for in the
> vocabulary.
> >
> > No, please do not blow up the foaf spec with properties and classes
> > for all individual services! I would prefer universally usable
> > concepts. We already have foaf:account and foaf:OnlineAccount :)
>
> Yes, definitely +1
>
> I would instead vote for deprecating service-specific properties and
> force the usage of universal ones (e.g. as the modelling example from
> Michael below).
> One can address a service separately in such a description, e.g., with
> the help of is:info_service [1] property of the Info Service Ontology
> (which is intended to describe and rate etc. (web) information services).
> Of course, utilising a direct binary relation to address that a specific
> ID is related to a specific service would simply on a first view the
> modelling somehow; however, this is not durable at all. Today Twitter is
> very popular. In 5 or 10 years we probably will use other social
> services. An ontology should be designed in such a way that it is
> durable (like the Web itself).
>
>
> >
> >> Not only is this extremely verbose to express a by now everyday data
> item, but I also have a problem when flattening it out to triples. I can
> only do this using blank nodes, which I always find ugly. Any known
> workaround for this?
> >
> > No, you could even use IDs or resource URIs:
> >
> > <foaf:Person rdfs:about="#you">
> >    <foaf:account rdfs:resource="#yourtwitter"/>
> > </foaf:Person>
> > <foaf:OnlineAccount rdfs:about="#yourtwitter">
> >    <!-- describe it here -->
> > </foaf:OnlineAccount>
>
> Cheers,
>
>
> Bob
>
>
>
> [1] http://purl.org/ontology/is/core#info_service
> _______________________________________________
> foaf-dev mailing list
> foaf-dev at lists.foaf-project.org
> http://lists.foaf-project.org/mailman/listinfo/foaf-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.foaf-project.org/pipermail/foaf-dev/attachments/20110617/2c7169c0/attachment.htm 


More information about the foaf-dev mailing list